Abstract
The Internet of Things (IoT) promises transformative advances in pervasive sensing, seamless connectivity, and intelligent computation. However, widespread deployment remains hindered by leakage currents, limited operational range, stringent power budgets, and performance bottlenecks. Because IoT nodes integrate diverse subsystems, including front-end energy harvesters, back-end regulators, data storage, and computation, isolated power optimization in a single block is insufficient to achieve true energy efficiency. This dissertation focuses on holistic, cross‑circuit design strategies that enable energy autonomy, extend operational range, and deliver robust performance across all functional domains of IoT devices. For the front-end circuits, we explored innovative structures of the circuits that combine adaptive body biasing with flipped‑well transistors and impedance‑matching optimization to realize RF–DC rectifiers with exceptional sensitivity under weak UHF signals. A comprehensive adaptive body‑biasing model is then developed to further maximize rectifier output voltage. In the back-end, a feedforward dynamic leakage suppression (FFDLS) two‑phase clock generator, together with active capacitor regulation, achieves ultra‑low‑power DC conversion. For data storage, a novel SRAM cell architecture is introduced that leverages the FFDLS technique to suppress static power while preserving speed and robustness across process, voltage, and temperature (PVT) variations. In computation, advanced FinFET technology is exploited to implement direct RF‑powered adiabatic logic (AL) circuits based on efficient charge recycling logic (ECRL), enabling high‑performance processing with reduced energy per cycle. Furthermore, positive feedback adiabatic logic (PFAL) is employed to design AC‑powered systolic array architectures for multiplication–accumulation (MAC) operations. The AL MAC systolic array is powered by a resonant 4-phase clock generator to sustain high clock rates (GHz level) with minimal power penalties. These contributions of this research establish a suite of low‑power, high‑sensitivity circuit building blocks that form a scalable foundation for next‑generation IoT devices, empowering them with energy autonomy, extended range, and resilient performance in diverse application environments.
